This article is part of a series on cricket analysis at club, age group performance and school level. To see part one, click here.
Win Percentage (Win%) is the R+ for the second innings in cricket. We can use the fluctuations in chances of victory to examine bowling and batting performance in a limited overs chase.
Most people are aware of Win% due to the WASP that appears in TV coverage of top-level cricket. But it’s more than entertainment. Coaches and players can use the metric to discover performance in context. For example, a team chasing starting at a Win% of 20% must work a lot harder than a team with a Win% of 80%. This works better than just the score as it takes the quality of the opposition and conditions into account.
So, at a team level, you can see the context of the challenge (is chasing 180 difficult or do you have an even chance?), then analyse how well you went about the chase after the match. It tells you how well you need to bat or bowl to win.
As an example, here is the average change in Win% during a season for a WDCU Premier Division team bowling second:
The chart shows the worst performance is in the first 10 overs, where there were five occasions of Win% falling, leading to a minus average. The middle overs see the team recover to between 9-13% further ahead. The last block is the biggest, as you would expect as the result is always clear at the end.
The clear lesson here is the team do let the opposition back into the game, but eventually chip away adding 1-2% Win% per over on average, showing the bowlers were doing a strong job adding Win%, even in games when the batsmen had made it easier.
At individual level, you can calculate how much each batting partnership and individual bowler and batsmen contributed to the win. This is called Win Percentage Added (Win%+)
If a bowler comes on when a team are at 50% Win% and bowls well enough to improve the Win% to 75%, you know they have done a strong job. However, Win% also deals with the context. The same bowler doing the same thing when Win% is at 25% will see the chances only get to 40%, nowhere near as effective despite being the same quality of performance. This is handy for analysis because you know who has exceeded expectations under pressure.
Top order batsmen, then, should average more than 9% Win%+ over a season chasing, unless the bowling unit is super-strong and they rarely need to chase an above Par score.
How to calculate Win%
Calculating Win% itself is simple, calculating Win%+ is more difficult because it requires over-by-over scores and takes longer to produce, so work with what you have.
Win% for the batting team is calculated:
Win% = 1/(1+(Par/Total)^8)
Where Par is the Par first innings score for the match and Total is the actual first innings score.
To work out the Win% for the bowling team, switch Par and Total around.
There is a challenge when breaking this down further to over-by-over updates in game, if you want live updates as in the televised matches. That challenge is wickets: If you are 100-2 after 20 overs chasing 200, you’re likely to win. If you’re 100-7, you’re unlikely to win.
Fortunately, there is an existing way of dealing with this; the DLS Par score. DLS takes wickets into account when calculating the total needed at the end of any over. So, create the table and apply Win% to it. Going back to our example, the DLS Par after 20 overs is 62 when two down and 156 when seven down. Win% is goes from 98% to 3%.
Using DLS you can update Win% every over; either live or in the post-game analysis to see where the fluctuations in momentum came.
Partnerships and Individual contributions: Win%+
Win%+ is, as we know, the amount of Win% an individual player – or batting partnership – contributes to the Win%. Working this out requires more work than Win%. Here are the steps.
For batting partnerships, first generate a DLS table for the Target score and calculate Win% at the end of each over. Win%+ is then:
Win%+ = W%o - W%i
Where Win%o is the Win% when the pair is broken up by a wicket, and Win%i is when the pair come together. This could be a positive or negative percentage.
For individual batsmen, the steps are similar. Generate the DLS table and Win% at each over. Then
Win%+ = (W%o - W%i) * %c
Where %c is the individual contribution as a percentage which is: Runs / Stand.
Finally, for bowlers, there are two ways to look at Win%+, simple and detailed. Detailed is significantly more accurate but is more effort to calculate (and requires DLS).
For the simple method, we apply the Win% calculation to the balls the bowler bowled. So if the opposition require 200 to win and the bowler bowls 10 overs, the Target is 40. If that bowler concedes 40 the Win% is 50% while they are bowling.
This method is quick and rather dirty but it does allow you to see which bowlers contained better than others more than pure economy rates.
The detailed method takes more context into account. Wickets taken and stage of the match (e.g. you are expected to concede more at the death than opening) alter the chances of victory.
Start by calculating DLS so you can know the Target at every over. For each bowling spell, do the following calculation:
Target = Pe – Ps / 2
Where Pe is DLS Par at the end of the spell and Ps is DLS Par at the start.
Combine these numbers for each spell to get a Target score, then perform the Win% calculation - Win% = 1/(1+(Par/Total)^8) – to get a Win%’.
To get the Win%+ you need two more things:
Team Win% at the start of the bowlers spell (TW%)
Number of balls bowled by the bowler as a percentage of total balls bowled in the game (B%)
Then the Win%+ calculation is:
Win%+ = (Win%’ - TW%) * B%
This could be a positive or negative percentage as bowlers who bowl poorly can give the opposition a better chance of winning just as much as good bowling reduces these chances.
While this approach takes more effort, it does tend to reveal more. For example, for West of Scotland CC’s bowling, the professional did not contribute the most Win%+. He was second to a player with a higher economy rate. If you simply looked at the averages, you may miss this and make tactical or training decisions based on out-of-context information.
Another example is in the Western Warriors bowling. Two spinners had almost identical analysis’s in two different games: 4-11 and 4-19. Which was the stronger contribution?
The 4-19 had a Win%+ of almost 9% while the 4-11 was just under 7%. Clearly both were excellent contributions, but the 4-19 had a greater impact on the outcome, so was more useful. We know from here that the 4-11 bowler is slightly better in a pressure situation when there are fewer runs to play with from the first innings. This is a very useful skill so is worth identifying!
Hopefully that is enough justification for getting your head into Excel. Next we will examine a way to simplify the numbers for players and turn it into one big number with the Impact score.